MRI Characterization of Pulsed Field Ablation Lesions in an Ex Vivo Porcine Model

Background: Pulsed field ablation (PFA) is a promising alternative to radiofrequency (RF) and cryoablation for treating cardiac arrhythmia. Unlike thermal RF ablation, PFA applies high-intensity electric (E)-fields to induce irreversible cellular electroporation, leading to subsequent cell death. PFA thus mitigates adverse thermal effects and is considered favorable due to its tissue selectivity. PFA lesions strongly depend on the pulse protocol and the E-field magnitude. Thus, controlling lesion formation remains challenging and profits from image guidance. MRI is the ideal imaging modality to characterize ablation lesions for both RF ablation and chronic PFA lesions. Thus, performing PFA already under MR guidance would enable the creation and characterizing of PFA lesions within a single intervention. In this study, the precision of MRI in assessing PFA lesion constitution is assessed in ex vivo porcine tissue, and the first MR-compatible system for PFA inside the MRI system is presented.

Methods: Six PFA lesions were created in the atria and the left ventricle of a female landrace pig (weight: 70kg, age: 3 months) using a focal catheter (Boston Scientific Intellanav Stablepoint) and a prototype RF generator (Stockert SmartRE PFA Generator C1.2). A monopolar protocol with biphasic asymmetric pulses up to 2kV was applied, and electroanatomical mapping was performed pre/post ablation (Boston Scientific Rhythmia HDx). Animals were euthanized 2h after ablation and 15min after intravenous injection of 0.5mmol/kg Gd contrast agent. In addition, 250ml of 1% triphenyltetrazolium chloride (TTC) was injected 10min prior to euthanasia. Both atria and part of the LV were fixated with formaldehyde and NaCl. Ex vivo MRI was performed 4 days post fixation at 3T (Siemens PrismaFit). T1- and R2*-weighted images were acquired at Δx=0.47mm isotropic resolution. Additionally, the devices were tested in the 3T MRI system was done using an MR-compatible monofocal catheter (Imricor Vision-MR) and an organic phantom.

Results: All lesions were visualized in both gross pathology and ex vivo MRI. In the T1-weighted images lesions appear hyperintense (contrast-to-noise ratio lesion/healthy tissue = 9.1±5.0). The mean lesion width/depth on ex vivo MRI were (9.2±3.8)/(5.2±1.4)mm compared to (8.3±3.4)/(4.6±2.0)mm on gross pathology, and a high linear correlation was found between MRI and pathology (R=0.98, p<0.001). The hemorrhagic parts of the lesions showed an increased R2* with a mean of (90.5±9.9)s^-1 compared to (23.2±1.0)s^-1 in healthy tissue, and their widths/depths were (5.7±1.9)/(3.8±2.0)mm on MRI and (5.5±1.8)mm (3.5±2.0)mm on gross pathology (R=0.99, p<0.001). The MR-compatible PFA system was successfully used to create electroporation in the organic phantom, validated by TCC staining.

Conclusion: This study demonstrates the feasibility of assessing PFA-induced lesions with MRI in an ex vivo porcine model. Lesion dimensions from MRI and gross pathology strongly correlate, and R2* values reliably identify the hemorrhagic lesion components. In combination with the MR-compatible PFA system, this method will be used for longitudinal studies on the formation of chronic lesions, providing valuable insights into PFA-based treatments of cardiac arrhythmia.